Start with real work
A useful assessment begins with everyday tasks, decisions and friction. It identifies where AI may shorten a delay, improve a decision or free up time—without confusing a demonstration with operational impact.
Assess four foundations
Business value, data availability and quality, governance (security, compliance and accountability), and industrialisation capacity must be assessed together. A critical weakness in any one area can prevent scaling.
Prioritise a testable roadmap
Compare use cases by expected value, feasibility, risks and prerequisites. The output is not a technology list: it is a sequence of experiments, decisions and foundations to build.
Validate before industrialising
A first use case can test business value and adoption conditions. Metrics, responsibilities and limits should be defined before production.